Current therapy for retinoblastoma utilizes surgical enucleation with resulting blindness and facial deformities or radiotherapy and chemotherapy that result in both local and systemic toxicity including an increase in the occurrence of second malignancies. Small tumors or large tumors that have been reduced in size by radiotherapy or chemotherapy can be treated successfully with laser photocoagulation or cryotherapy. Numerous preclinical and clinical trials have shown that adenoviral vectors used to deliver the Herpes thymidine kinase gene followed by ganciclovir can be safely administered. However, this approach has not been shown to be efficacious to cure large, aggressive malignancies such as brain tumors and prostate cancer. This phase I study is designed to show that this gene therapy approach can be used as an adjuvant treatment for children with retinoblastoma, a unique confined small tumor with relatively low malignant potential. The ultimate goal of this gene therapy is to salvage the affected eye by reducing the tumor burden sufficiently to allow for local control of the disease. Based on demonstrated potential for efficacy in preclinical laboratory studies, and in a preliminary clinical pilot study conducted at this institution, patients with nonmetastatic retinoblastoma are treated with a direct injection of an adenoviral vector that delivers the Herpes simplex thymidine kinase gene. Vector is delivered using cryotherapy before and after injection to minimize the chance of seeding the tumor. Ganciclovir is then administered intravenously at a dose of 10 mg/kg/day for 7 days. Up to a total of 4 injections of the viral vector will be considered depending on patient response. The objective of this phase I study is to determine whether this treatment is associated with significant toxicity. A secondary goal is to assess the efficacy of the therapy. If clinical efficacy is achieved before toxicity, the trial will be terminated at that dose of viral vector. The trial is designed as an intra-patient dose escalation study. The dose of the vector will begin at 1 x 1010 viral particles and, for the second dose, will increase by one-log increment in the absence of toxicity and progressive disease. Three patients will be treated using this treatment strategy and if no toxicity is found the dose will be increased by one log increment and begin at 1 x 1011 viral particles. This research is the first step towards our ultimate goal, which is to develop a new generation of more effective, less toxic therapies for patients with ocular diseases